PDA

View Full Version : Mirror paradox?



94z07
2006-Jul-27, 01:47 PM
Is it possible to shine a flashlight on a mirror and reflect the light back to the flashlight such that light leaving the filament at point A travels to point B at the mirror and is then reflected back to point A on the filament?

If so does that mean that the light, when created, has an instantaneous velocity of C then slows to the speed of light in glass as it passes through the bulb then accelerates to the speed of light in air then slows to the speed of light in glass as it passes through the lens of the flashlight then accelerates to the speed of light in air until it reaches the mirrorís glass where it slows again to the speed of light in glass until it hits the aluminum where it actually stops?

If the light travels the same path from A to B that it travels from B to A it would seem that the light actually has an instantaneous velocity of zero at point B. Where it then instantaneously accelerates to the speed of light in glass as it leaves point B for the return trip to A.

----- OR -----

Does the light never actually stop because it never actually reaches point B? Does the photon perform some tight orbit of an aluminum atom and head back to point A?

ToSeek
2006-Jul-27, 02:10 PM
I can't find a really good explanation online, but as near as I can figure the photon is actually absorbed by an electron at point B, which goes into a more excited state until it re-emits an identical photon in the appropriate direction. (Why the electron emits the photon in just the right direction requires a better understanding of solid-state physics than I've got.)

Sticks
2006-Jul-27, 02:12 PM
You also need to account for coherence and polarity.

papageno
2006-Jul-27, 03:43 PM
I can't find a really good explanation online, but as near as I can figure the photon is actually absorbed by an electron at point B, which goes into a more excited state until it re-emits an identical photon in the appropriate direction.
I think you can do it more classically.
When the wavelength of the light is much larger than the typical size of an atom (light ~ 500 nm, atom ~ 1 nm), we can imagine the atom as an electric dipole: the electric field of the EM wave separates the barycentre of the negative charge and the positive charge, polarizing temporarily the atom. The atom then behaves as an oscillating dipole, producing an EM wave, whose maximum amplitude is in directions parallel to the dipole axis, including the direction where the original EM wave came from.
If you have lots of atoms closely packed together, each emitting EM waves, you have interference of lots of tiny spherical EM waves.

If we are talking about a mirror, this is not the whole story.
The reflective part of a mirror is typically a metallic layer.
Part of the electrons in a metal are not bound to specific atoms, but can move freely inside the whole metal. These conduction electrons form a gas which behaves as a plasma. One of the characteristics is a crtical frequency: EM waves with frequency lower than the critical frequency cannot propagate inside the metal (because it cannot excite the system into an allowed state) and is reflected.



(Why the electron emits the photon in just the right direction requires a better understanding of solid-state physics than I've got.) One one hand there are the conservation laws for the EM interaction (and the so-called selection rules restricting the possible electronic transitions in atoms and molecules), on the other hand you have interference from lots of atoms (Huygens principle in optics, if I remember correctly).

umop ap!sdn
2006-Jul-27, 06:04 PM
These conduction electrons form a gas which behaves as a plasma. One of the characteristics is a crtical frequency: EM waves with frequency lower than the critical frequency cannot propagate inside the metal (because it cannot excite the system into an allowed state) and is reflected.
Interesting! Makes sense as metals' ability to block X-rays is IIRC due to absorption by the nucleus. I wonder then what wavelength the cutoff is; do metals suddenly become transparent somewhere in the vacuum ultraviolet or EUV?

antoniseb
2006-Jul-27, 06:28 PM
the wavelength of the light is much larger than the typical size of an atom (light ~ 500 nm, atom ~ 1 nm)

Nice explanation, though I'd like to amend it to say that the typical size of an atom is 0.1 nm.

I also like to make the analogy that the fields from the photon interact with the fields from the electrons in the reflector collectively and that they give and restore themselves, and that restoration can be thought ot as causing the return photon.

It should also be noted that there is no change in momentum. The photon does not carry mass, so there is no paradox.

papageno
2006-Jul-27, 06:53 PM
Interesting! Makes sense as metals' ability to block X-rays is IIRC due to absorption by the nucleus. I wonder then what wavelength the cutoff is; do metals suddenly become transparent somewhere in the vacuum ultraviolet or EUV?
The dielectric constant of the electron gas in a conductor, as a first approximation*, takes the form:

epsilon(omega) = 1 - (omega_p/omega)^2

where omega is the frequency and omega_p the plasma frequency (which depends on electron density and effective mass). Usually the formula is derived from the AC conductivity of a metal.

When omega < omega_p then epsilon < 0, and EM waves cannot propagate in the metal.
When omega > omega_p EM waves can propagate.
So, omega_p could be considered as a cut-off frequency, though the increase of epsilon is continuous from zero.

Using excitations (shooting photons or electrons at the metal) with energies that are integer multiples of hbar omega_p, it is possible to excite plasmons, that is quantized oscillation in the density of conduction electrons. (We did that in a teaching lab, where we excited a handful of plasmon in a piece of Al using electrons.)



* Approximation: being far from funny resonances and the average scattering rate of an electron much higher than the frequencies involved.

papageno
2006-Jul-27, 06:59 PM
Nice explanation, though I'd like to amend it to say that the typical size of an atom is 0.1 nm.
:doh: I had 1 Angstrom = 0.1 nm in mind.




It should also be noted that there is no change in momentum. The photon does not carry mass, so there is no paradox.
But a photon does carry momentum p = E/c , so in principle one should take into account conservation of momentum.
However, the momentum of EM waves is typically too small to yield significant effects in these situations.

94z07
2006-Jul-28, 02:57 AM
Thank you all for taking the time to answer my question so well.

Stick's reply "...coherence and polarity" got me into a couple hour's reading and googling.

Maybe in the future I'll format my questions to yield such replies.